NBER Digest reporter Claire Brunel covers a new paper by Christopher Knittel and Ryan Sandler on gas taxes and air pollution that makes an interesting point (see the end of the post in bold) about uniform gas taxes.
While some have questioned the benefits of mass transit systems, which are used by only a small fraction of commuters, research by Michael Anderson suggests that transit riders likely would otherwise commute along already heavily
congested roadways — and that congestion along those roadways would increase if mass transit were scaled back. In Subways, Strikes, and Slowdowns: The Impacts
of Public Transit on Traffic (NBER Working Paper No. 18849), he studies traffic-congestion data from before, during, and after an October-November 2003 transit-worker strike in Los Angeles. He estimates that average highway congestion delays increased 47 percent when public transit service was not
Driving a car produces an externality — pollution — that imposes various economic costs, but drivers do not directly pay for that externality. The efficient economic remedy would tax individual drivers based on their vehicles’ emissions, but that is currently impractical. Taxing gasoline purchases, which
are related to emissions, is an indirect alternative, but the gasoline tax does not take into account the variation across vehicles in per-gallon emissions. In The Welfare Impact of Indirect Pigouvian Taxation: Evidence from Transportation
(NBER Working Paper No. 18849), Christopher Knittel and Ryan Sandler examine how differences in consumers’ price sensitivity of demand for gasoline, and variation in per-gallon emissions across vehicles, affect the efficiency of the
gasoline tax as a policy instrument for reducing pollution.
Using vehicle data from California for the period 1998 to 2008, they find that the variation in vehicle emissions is correlated with how vehicle-specific-miles-driven respond to a change in gasoline prices. The drivers of dirtier vehicles respond substantially more to changes in fuel prices
than the drivers of clean vehicles. This means that the local pollution-lowering benefits from an increase in the gasoline tax are larger than estimates based on equal-demand responses across all drivers. In fact, the authors estimate that the health benefits of a gasoline tax would increase by 90 percent once the variation in responsiveness of vehicle emissions
is taken into account. This result is not driven by a vintage effect, whereby older vehicles are more responsive to changes in gasoline prices and at the same time have higher emissions.
The authors find that a uniform tax on emissions — currently impractical but often discussed — would perform poorly in eliminating the inefficiencies
associated with vehicle emissions. Over 75 percent of those inefficiencies would remain with such a tax. The dirtiest vehicles, which are most responsive to an emissions tax, would not be taxed enough while some clean vehicles would be over-taxed. One potential strategy for reducing these inefficiencies would be to condition taxes based on vehicle type, and particularly on the age of the vehicle. Another route would be to scrap the dirtiest 10 percent of vehicles, or to retire the most polluting vehicles through cash-for-clunkers or allowing gasoline taxes to be county-specific.